Math & Logic
Math & Logic8 min read

Bayes' Theorem in Plain English — The Most Useful Math You Never Learned

Published May 16, 2026

The Problem with Human Reasoning

We're bad at updating our beliefs. When new evidence comes in, we tend to either ignore it or overreact to it. We anchor on first impressions. We seek confirmation.

Bayes' theorem is a mathematical formula for doing this correctly. It's used in spam filters, medical diagnosis, machine learning, and intelligence analysis.

The Core Idea

Bayes' theorem tells you how to update the probability of a belief given new evidence.

In plain English: How probable is this belief, given what I now know?

The formula is: P(A|B) = P(B|A) × P(A) / P(B)

Don't panic. Let's use an example.

The Medical Test Problem

Imagine a disease affects 1% of the population. A test for this disease is 99% accurate (it correctly identifies sick people 99% of the time, and correctly clears healthy people 99% of the time).

You test positive. What's the probability you actually have the disease?

Most people say 99%. They're very wrong.

Working through it:

  • In a population of 10,000: 100 have the disease, 9,900 don't
  • The test catches 99 of the 100 sick people ✓
  • But the test also wrongly flags 1% of the 9,900 healthy people = 99 false positives

So you have: 99 true positives + 99 false positives = 198 total positive tests Your chance of actually being sick: 99/198 = 50%

A 99% accurate test gives you a coin flip result — because the disease is rare. This is why doctors don't test everyone for everything, and why base rates matter enormously.

The Practical Application: Belief Updating

You don't need to run the formula in daily life. The habit it teaches is the valuable part:

  1. Start with a prior probability. What do you currently believe, and how strongly?
  2. Assess the evidence quality. How reliable is this new information? What's the source?
  3. Update proportionally. Strong evidence from a reliable source moves you more than weak evidence from an unreliable one.
  4. Acknowledge uncertainty. You're always working with probabilities, not certainties.

The Meta-Lesson

Bayesian thinking is fundamentally humble. It says: you have beliefs, but they're probabilistic. New information should move them — but not completely, and not without weighing the quality of that information.

It's the mathematical formalization of "keeping an open mind" — with the crucial addition that how open your mind should be depends on the evidence.